Turn Clean Data into Messy Data

Take real or simulated data and salt it with errors commonly found in the wild, such as pseudo-OCR errors, Unicode problems, numeric fields with nonsensical punctuation, bad dates, etc.


lifecycle Travis-CI BuildStatus AppVeyor BuildStatus CoverageStatus

When teaching students how to clean data, it helps to have data that isn’t too clean already. salty offers functions for “salting” clean data with problems often found in datasets in the wild, such as:

  • pseudo-OCR errors
  • inconsistent capitalization and spelling
  • invalid dates
  • unpredictable punctuation in numeric fields
  • missing values or empty strings

Installation

You can install salty from github with:

# install.packages("devtools")
devtools::install_github("mdlincoln/salty")

Basic usage

library(salty)
set.seed(10)
 
# We'll use charaltan to create some sample data
 
sample_names <- charlatan::ch_name(10)
sample_names
#>  [1] "Edwin Kassulke"       "Barron Fadel"         "Dorla Morissette"    
#>  [4] "Manuela Mante MD"     "Ferris Kautzer"       "Djuana Hyatt"        
#>  [7] "Dr. Leighton Ryan"    "Ms. Migdalia Smitham" "Ottilia Hermann"     
#> [10] "Benjiman Dach"
 
sample_numbers <- charlatan::ch_double(10)
sample_numbers
#>  [1]  1.280597456  0.667415054  1.691754965  0.001261409 -0.742461312
#>  [6]  0.609684421 -0.989606379 -0.034848335  0.847159906  1.525498006

salty offers several easy-to-use functions for adding common problems to your data.

# Add in erroeous letters or puncutation
salt_letters(sample_names)
#>  [1] "Edwin Kassulke"       "Barroun Fadel"        "Dorla Morissette"    
#>  [4] "Manuela Mante MD"     "Ferris Kyautzer"      "Djuana Hyatt"        
#>  [7] "Dr. Leighton Ryan"    "Ms. Migdalia Smitham" "Ottilia Hermann"     
#> [10] "Benjiman Dach"
salt_punctuation(sample_names)
#>  [1] "Edwin K'assulke"      "Barron Fadel"         "Dorla Morissette"    
#>  [4] "Manuela Mante MD"     "Ferris Kautzer"       "D$juana Hyatt"       
#>  [7] "Dr. Leighton Ryan"    "Ms. Migdalia Smitham" "Ottilia Hermann"     
#> [10] "Benjiman Dach"
 
# Flip capitals
salt_capitalization(sample_names)
#>  [1] "Edwin Kassulke"       "Barron Fadel"         "Dorla Morissette"    
#>  [4] "Manuela Mante MD"     "Ferris Kautzer"       "Djuana Hyatt"        
#>  [7] "Dr. Leighton Ryan"    "Ms. MigdalIa SmItHam" "Ottilia Hermann"     
#> [10] "Benjiman Dach"
 
# Introduce OCR errors. You can specify the proportion of values in the vector
# that should be salted, and the proportion of possible matches within a single
# value that should be changed.
salt_ocr(sample_names, p = 1, rep_p = 1)
#>  [1] "Edwvi'n Kassulke"      "BarroIn Fadel"        
#>  [3] "Dorla Morisfette"      "Mlanuela Mlante MD"   
#>  [5] "Ferris Kautzer"        "Djuana Hyatt"         
#>  [7] "Dr. LeiglhltoIn Ryan"  "Ms. Migdalia Smitlham"
#>  [9] "Ottilia Hermann"       "Benjiman Daclh"

salt_delete will simply drop characters from randomly selected values in a vector, while salt_empty and salt_na will replace entire values.

salt_delete(sample_names, p = 0.5, n = 6)
#>  [1] "Edwin Kassulke"   "Barron Fadel"     "Dor Morset"      
#>  [4] "Manuela Mante MD" "Feri Kauz"        "Djuana Hyatt"    
#>  [7] "r. Lightoan"      "MsMidala Smiha"   "OttliaHean"      
#> [10] "Benjiman Dach"
 
salt_empty(sample_names, p = 0.5)
#>  [1] ""                     ""                     "Dorla Morissette"    
#>  [4] "Manuela Mante MD"     ""                     ""                    
#>  [7] "Dr. Leighton Ryan"    "Ms. Migdalia Smitham" "Ottilia Hermann"     
#> [10] ""
 
salt_na(sample_names, p = 0.5)
#>  [1] "Edwin Kassulke"       NA                     NA                    
#>  [4] NA                     "Ferris Kautzer"       "Djuana Hyatt"        
#>  [7] NA                     "Ms. Migdalia Smitham" "Ottilia Hermann"     
#> [10] NA

Advanced usage

For more fine-grained control over the salting process, and for access to a wider range of salting types, you can use the underlying functions provided for: inserting, substituting, replacing.

The set of insertions and replacements are specified via shakers, pre-filled character sets and pattern/replacement pairs that the salt verbs then call.

available_shakers()
#> $shaker
#> [1] "punctuation"       "lowercase_letters" "uppercase_letters"
#> [4] "mixed_letters"     "whitespace"        "digits"           
#> 
#> $replacement_shaker
#> [1] "ocr_errors"     "capitalization" "decimal_commas"

salt_insert keeps all the characters in the original while adding new ones, while salt_substitute overwrites those characters.

# Use p to specify the percent of values that you would like to salt
salt_insert(sample_names, shaker$punctuation, p = 0.5)
#>  [1] "Ed\"win Kassulke"      "B^arron Fadel"        
#>  [3] "Dorla Morissette"      "Manuela Mante MD"     
#>  [5] "Ferris Kautzer"        "Djuana Hyatt"         
#>  [7] "Dr. Leighton Ryan"     "Ms.( Migdalia Smitham"
#>  [9] "Ottil.ia Hermann"      "Benj$iman Dach"
 
# Use n to specify how many new insertions/substitutions you want to make to selected values
salt_substitute(sample_names, shaker$punctuation, p = 0.5, n = 3)
#>  [1] "Edwin Kassulke"       "Barron Fadel"         "D/rla Mo.issette."   
#>  [4] "Manuela Mante MD"     "Ferris %a^t*er"       "Dju,na^Hyatt'"       
#>  [7] "Dr. Leighto\" *(an"   "Ms. Migdalia Smitham" "O%tili^ [email protected]"    
#> [10] "Benjiman Dach"

Different flavors of salt are available using shaker, however you can also supply your own character vector of possible replacements if you like.

salt_insert(sample_names, shaker$mixed_letters, p = 0.5)
#>  [1] "Edwin Kassulke"       "Barron FLadel"        "Dorla Morissette"    
#>  [4] "Manuela MantIe MD"    "Ferris Kautzer"       "Djuana Hyatt"        
#>  [7] "DrU. Leighton Ryan"   "Ms. Migdalia Smitham" "Ottilia Hermannn"    
#> [10] "Benjiman DachM"
 
salt_insert(sample_numbers, shaker$digits, p = 0.5)
#>  [1] "1.328059745613008"    "0.667415054241444"    "1.69175496457426"    
#>  [4] "0.001261408793618831" "-0.7424613118147763"  "0.6096844205304159"  
#>  [7] "-20.989606379077806"  "-0.0348483349098612"  "0.847159905848433"   
#> [10] "1.52549800647527"
 
salt_insert(sample_names, c("foo", "bar", "baz"), p = 0.5)
#>  [1] "Edwin Kassulke"       "Barron Fadel"         "Dorla Morissette"    
#>  [4] "Manuela Mantebaz MD"  "Ferrfoois Kautzer"    "Djuanabar Hyatt"     
#>  [7] "Dr. Leighton Ryan"    "Ms. Migdalia Smitham" "Ottbazilia Hermann"  
#> [10] "Benjiman Dacbarh"

salt_replace is a bit more targeted: it works with pairs of patterns and replacements, either contained in replacement_shaker or user-specified. Use rep_p to set a probability of how many possible replacements should actually get swapped out for any given value.

salt_replace(sample_names, replacement_shaker$ocr_errors, p = 1, rep_p = 1)
#>  [1] "Edwvi'n Kassulke"      "BarroIn Fadel"        
#>  [3] "Dorla Morisfette"      "Mlanuela Mlante MD"   
#>  [5] "Ferris Kautzer"        "Djuana Hyatt"         
#>  [7] "Dr. LeiglhltoIn Ryan"  "Ms. Migdalia Smitlham"
#>  [9] "Ottilia Hermann"       "Benjiman Daclh"
 
salt_replace(sample_names, replacement_shaker$capitalization, p = 0.5, rep_p = 0.2)
#>  [1] "Edwin KassUlKe"       "bARRon FaDeL"         "Dorla Morissette"    
#>  [4] "MAnuelA MAnTe MD"     "fErris KautZer"       "Djuana Hyatt"        
#>  [7] "Dr. Leighton Ryan"    "Ms. Migdalia Smitham" "Ottilia Hermann"     
#> [10] "Benjiman Dach"
 
salt_replace(sample_numbers, replacement_shaker$decimal_commas, p = 0.5, rep_p = 1)
#>  [1] "1,28059745613008"    "0.667415054241444"   "1.69175496457426"   
#>  [4] "0.00126140879361831" "-0,742461311814763"  "0,609684420504159"  
#>  [7] "-0,989606379077806"  "-0.0348483349098612" "0.847159905848433"  
#> [10] "1,52549800647527"

You may also specify your own arbitrary character vector of possible insertions.

salt_insert(sample_names, insertions = c("X", "Z"))
#>  [1] "Edwin Kassulke"       "Barron FadZel"        "Dorla Morissette"    
#>  [4] "Manuela Mante MD"     "Ferris Kautzer"       "Djuana HyatXt"       
#>  [7] "Dr. Leighton Ryan"    "Ms. Migdalia Smitham" "Ottilia Hermann"     
#> [10] "Benjiman Dach"

Possible future work

  • Modifying date strings to introduce subtle errors like invalid dates (e.g. February 30th)
  • Simulting character encoding problems

Related work

salty should not be used for anonymizing data; that’s not its purpose. However, it does draw some inspiration from anonymizer.

To create sample data for salting, take a look at charlatan.

Acknowledgements

The common OCR replacement errors are partially derived from the sed replacements specified in the Royal Society Corpus project: Knappen, Jörg, Fischer, Stefan, Kermes, Hannah, Teich, Elke, and Fankhauser, Peter. 2017. “The Making of the Royal Society Corpus.” In Proceedings of the NoDaLiDa 2017 Workshop on Processing Historical Language. Göteborg, Sweden. Linköping University Electronic Press. http://www.ep.liu.se/ecp/article.asp?issue=133&article=003&volume=.

News

salty 0.1.0

  • First CRAN release of salty
  • Added a NEWS.md file to track changes to the package.

Reference manual

It appears you don't have a PDF plugin for this browser. You can click here to download the reference manual.

install.packages("salty")

0.1.0 by Matthew Lincoln, a year ago


https://github.com/mdlincoln/salty


Report a bug at https://github.com/mdlincoln/salty/issues


Browse source code at https://github.com/cran/salty


Authors: Matthew Lincoln [aut, cre]


Documentation:   PDF Manual  


MIT + file LICENSE license


Imports assertthat, purrr, stringr

Suggests charlatan, testthat, tibble, covr


See at CRAN